Gray Gradients can't be applied to a & b Channels of Lab?

[Koyaanis]

I was using the information contained on this page "How to Create Color Test Charts" on how to create your own test-charts for printing and color-space management. Using the information on that page, making the Granger Chart worked just fine in Photoline. But when it comes to making the Gamut chart, I couldn't get it to work. If I try to apply a gray gradient to either the a or b channels in a blank white Lab image, nothing happens. Yet when I try to apply any color gradients to those channels, then I get the intended effects.

I'm not well versed in using Lab color channels, so I presume I'm doing something wrong, or some extra property of that background or layer must be enabled to make this work.

So ... using the info on that page, how do I make a Gamut Chart?

It says to (using PS commands no doubt):

Create a new file. CTRL / CMD – N. I make mine 10 inches by 10 inches at 300 ppi in 16 bit mode. Make sure that the Background colour is White.
Convert the file to LAB mode; Image / Mode / Lab Color.
Select the Channels palette. Select the 'a' channel.
Select the gradient tool (G) and then the 'Black, White' gradient. Set mode to 'Normal' and 'Opacity' to 100%. Draw a gradient from left to right across your document while holding down the Shift
key. (This keeps the line and the gradient straight).
Select the 'b' channel.
Using the current gradient settings draw a gradient from the bottom to the top of your document while holding down the Shift key.
Click on the Lab channel.
Convert the file back to RGB mode; Image / Mode / RGB Color.
Save the file in a place where you won't misplace it. I call it Gamut Chart.

To get this:

[Hoogo]

Yes, it looks to me that a single active channel does not mean that the brightness is stored to that channel. Instead the brush-color keeps being a brush-color. A bw-gradient goes from Lab 0/50/50% to 100/50/50%, so painting with a single activated a or b-channel sets the channels to 50%.
I can think of 2 ways to do that:
-Draw a L-gradient and use "tool > change channels".
-Imho easier: Select the gradient-tool, double-click on the starting color to open the color-editor, there you change to Lab and choose the desired a/b-color. Same with the end-color, then choose the channel and draw the gradient.

[Koyaanis]

Thanks, but I'm not sure what colors to select to make that Lab image mathematically correct. I'll play around, see what I can come up with (but then never be certain I did it right).

Interestingly, when looking for a ProPhoto color profile to install on my computer (what lead up to this whole test-chart thing, mentioned on this page Understanding ProPhoto RGB), I ran across this add-on from MickySoft, a color-profile control panel for free that you can install on XP, looks interesting.

time passes ...

Well, I finally got it to work. None of the methods you suggested worked. But a variation of one did. I made 2 L-gradients in the required directions in a Lab document, then saved them as Photoline documents (*.pld). Then using the Change Channels tool, I used that to move the files into the required channels.

Seems like quite the work-around for something that should work. (btw: PL gives a "major error" warning if you try to open a file in the Change Channels tool but don't actually select one to import into a channel).

After seeing what had to be done to get it to work, methinks that PL32 needs some work in this area.

Hmm... interesting. I accidently clicked on my Granger Chart that I made with PL32, while I had the flood-fill tool still active with a gray gradient. The tool's settings were set at Tolerance 100%, but it only filled in an area of color. I thought tolerance 100% meant that it would match ANY color or luminosity. Yet another area that PL32 needs fixing?

And one more thing that needs fixing. My "UNDO" sometimes doesn't undo to the beginning. It sets the first step as sometimes the 1st, 2nd, or 3rd step. Maybe this is a memory limitation or something. Working with these large 16 bit files might be showing me a limit I didn't know I had. (Duh, that was it, I had left PL32's default Undo set to 128MB.)

Oh, one more weirdness that I noticed while doing this. When flood-filling a gradient a second time, the gradient would be off-set to one side or the other, filling up only part of my image. I never could figure out what was causing that, no matter how much I tried to center the mouse before flood-filling, it would shift it off to the side and leave part of the image untouched.

(While testing all this, 6 deer were grazing right outside the house, and are now laying down next to the house while I type this, so I had to snap off some photos of them too. )

[Martin Huber]

Thanks, but I'm not sure what colors to select to make that Lab image mathematically correct. I'll play around, see what I can come up with (but then never be certain I did it right).

You want an image, that covers the whole range of a and the whole range of b. Given, that you already have created the white Lab-image, the easiest way to do this in PhotoLine is:
- Select the blend tool
- Define a blend with 2 colors. The start and end color have to be in Lab-mode.
The start color should be L: 100%, a: 0% b: 0%
The end color should be: L: 100%, a: 100%, b: 100%
- In the tool settings of the blend turn off L and b and turn on a (if it isn't already on).
- Draw the horizontal blend.
- In the tool settings turn off a and turn on b.
- Draw the vertical blend.

This should produce the gamut chart. Depending on the direction of the blends, the gamut chart may be flipped.
I think this way is easier than the work-around using Change Channels.

Martin

[Koyaanis]

Thanks for the tips and method. As I said, not having played much in the Lab channels before, this was confusing me.

This still leaves me wondering why the "Flood Fill" tool with a 100% Tolerance settings will only match a small range of colors in these charts. Is it because these are in 16-bit mode and the 100% tolerance setting can't "see" all these extra colors? If so, I think that needs to be changed. No?

[Koyaanis]

Hmm... it still doesn't work. When I add a blend to the "b" channel then it only fills up part of the image, probably due to that Intensity value not matching all values again.

I guess I'll just save a blend and then import it to a channel, at least that worked. (This really needs to be fixed, it's broken on two levels. Can't apply values to Lab channels, and can't apply flood-fills to a full image due to the color matching preventing it.)

[lutz]

.....
This still leaves me wondering why the "Flood Fill" tool with a 100% Tolerance settings will only match a small range of colors in these charts.

I have also encountered this problem several times - even with 8 bit images. What could we be doing wrong here? Why does a 100% tolerance setting not fill everything?